Socially Assistive Robot in Upper-Limb Neurorehabilitation

Randomized Control Trial (RCT) longitudinal Intervention study. The RCT will be in accordance with the Consolidated Standards of Reporting Trials (CONSORT) statement . In this study there will be 2 groups of intervention and one control group:

1. Socially Assistive Robotic (SAR) intervention group.
2. Computer intervention group.
3. Control group: standard therapy. Participants A total of 30 subacute, one year post first stroke survivors, age 45-85, females and males, will be randomly allocated into 3 intervention groups. The participants will be recruited from Aleh Negev rehabilitation center, from the ambulatory day care by Dr. Oren Barzel-MD, a Rehabilitation Medicine specialist.

Procedure All patients that will be recruited to the study will be randomly allocated to each of the three intervention groups. The measurements and therapy sessions will take place in Aleh Negev rehabilitation center, in the Day-Care unit, by a physiotherapist and a PhD student from Ben Gurion University (GCP).

Evaluations for all patients will be conducted in three points of time: T1: entrance to the study, before randomization. T2: following 15 intervention sessions (after 5 weeks). T3: one month after end of the intervention. MMSE and FM-UE scores will be recorded as screening tools. FM-UE is a screening tool and a primary outcome measures. At each time point (T1, T2 and T3), the measurements described below will be performed on two separate days to avoid effect of fatigue. Participants will come for a total of 21 sessions: 2 sessions of evaluations, 15 sessions of treatment, 2 sessions of evaluation at the end of the intervention, 2 sessions of evaluations one month post intervention.

Intervention Program The intervention program consists of 15 therapy sessions, 3 times a week, an hour session each time for 5 weeks (over all 15 hours of intervention). Our proposed intervention is a graduated functional RTG games using real objects (like a cup or a jar), in different sizes, shapes, weights, textures and different environments (i.e. different heights or location on the table). The RTG game selected for a participant will be based on the upper limb movement and functional ability of the participant, as was measured by the therapist in the inclusion to the study. The progression will be starting from low table and progressing to higher table, starting from low weights objects (i.e. an empty jar) and progressing to higher weights (i. e. a jar full with flour). Our program is based on two critical parameters of upper limb rehabilitation: 1. the amount of practice-it has been recommended that individuals post-stroke perform as many repetitions per session as they can tolerate . 2. Practicing challenging tasks and not simple repetitive tasks, is likely to elicit motor learning and associated neural reorganization. Since a correlation between cognition and arm motor improvement has been lately described, the games we developed contain upper limb movement as well as cognitive aspects (like working memory). In each meeting the participant will have a one-hour session of RTG games, based on his\ her ability, as was measured upon enrollment in the study and based on his\her fatigue etc. The patient will be able to rest up to 5 minutes between a trial to trial in order to avoid fatigue. In addition, a patient will be able to stop a session at any time if desired. Even though our aim is to evaluate the motivation and performance of the patient outside the individual physiotherapy session with the therapist, we assume that during self-practice at the clinic and at home there is someone with the patient. Therefore during the therapy session of the intervention study a physiotherapist or a student will be present in the room (to help with problems that can arise). All sessions will be filmed and recorded upon agreement of the participant.

Group 1: Upper-limb functional RTG games with a humanoid robot (the "ROBOT group"). The robot that will be used in this study is the Pepper robot (Aldebaran Soft Bank robotics) -a humanoid off the shelf robot- with human-like face, 120 cm in height (see Figure 1). Pepper has a touch screen tablet on its chest. The participant will sit in front of the robot, and will play RTG games with the robot using real objects like a cup or a jar, open a drawer, order jars on a shelf etc. The robot will give the participant feedback on his\ her performance: the speed and the accuracy of the movement. There is no physical touch between the robot and he participant. There are no modulation of the off-the-shelf features of the robot for the study, besides recording the sentences of the robot speech for Hebrew in order for the patients to understand his speech. Pepper is off-the shelf humanoid robot, with human-like face. It can interact with the patient by gestures of voice (talk to the patient, encourage him\her to keep on training), by gestures of movement (like hand clapping, hand raising etc.) and by changing its color of eyes.

Group 2: Upper-limb functional RTG games with a computer (the "COMPUTER group"). The games are the same as described for the "ROBOT" group. In this group the pictures of the cups and the feedback on performance will be displayed on a standard computer screen (17''). There is no physical touch between computer and the participant.

Group 3: follow up with no intervention besides the usual rehabilitation program at the facility.

An example of a game: the cup game. In each trial, a row of colored cups will be displayed on the robot's chest tablet. The participant have to organize a corresponding set of actual physical cups on the table according to the picture shown on the screen. There are three levels of game difficulty: in Level 1 participants are asked to arrange four cups in the correct order, in Level 2 there are five cups, and in Level 3 there are six cups. The instructions and feedback will be provided by either the robot, or displayed on the computer screen or written on a paper. The picture with the target order of the cups will disappear from the screen after 1-2 min. In the computer and robot groups, when they complete the task, participants will have to press a push button. In the robot and computer groups, after each trial, the robot\ computer will either give feedback on the performance (e.g., "try to do it faster next time") or a feedback on results (e.g., "you succeeded!", "you were not right but try again!"). The feedback given by the robot is both verbal and with head and arm gestures (e.g. hand clapping). The feedback by the computer will be displayed on the screen. In the self-training group no feedback is given. After two trials on the basic level (four cups) and four trials on the second level (five cups), the participant could choose to either continue at the same level or change to a harder level of game (more cups).

1. Games description and progression
2. The cup game: Ordering a row of cups according to the picture. Progression 1: number of cups: start from 3 cups, continue to 6 cups. Progression 2: Height of table: start low, continue to high, continue to different heights, different weights of cups.
3. A dartboard: ordering cups on a dartboard according to a picture. Progression 1: number of cups: start from 3 cups continue up to 7 cups. Progression 2: start close to body (at arm's length) continue longer than arm length. Progression 3: height of table. Progression 4: weight of cups.
4. Ordering Jars on different heights shelves: according to a picture. Progression 1: start 3 jars, add jars. Progression 2: start low height, continue high and different heights. Progression 3: different weights of jars.
5. Open drawer and order objects according to a picture (memory game). Progression 1: number of drawers-start with 2 and add drawers. Progression 2: different heights drawers. Progression 3: start with a handle of different kinds (straight handle, round handle) continue to a key opening.
6. Taking objects (i.e. keys) out of a jar and putting it in a jar according to its color (stroop game).